机器学习桥接城市内涝与气候风险:研究进展与展望OA

Climate change and urbanization have led to frequent extreme weather events and severe urban waterlogging risks,while conventional methods fail to bridge the gap between climate science and urban waterlogging management.Through systematic review and critical analysis,this paper for the first time summarizes three core"bridging"modes of machine learning connecting the two domains:multi-scale data assimilation and feature transformation bridge,model-ing bridge integrating physical mechanisms and data-driven approaches,and assessment bridge for uncertainty quantific-ation and risk transmission.This paper first analyzes the coupling mechanism between climate science and urban water-logging research and key challenges,then reviews key technologies supporting these bridging modes,and elaborates how each mode addresses core issues through case studies.This summarized framework provides a novel perspective for understanding and mitigating urban waterlogging risks under climate change.